SA Forum is an invited essay from experts on topical issues in science and technology.

Need to trade your favorite stock while sitting at home in your pajamas? No problem. High speed, point-to-point communications is something brokerage firms have long taken advantage of. Trading stock a split second faster than your competitor can translate into huge profits. But need to send images from a CT scan to Dallas from Detroit? The process will be less efficient and may well result in a repeat scan. In 2014 the average physician still processes over 1,100 faxes each month.

The contrast between banking and medicine underscores the sorry state of data connectivity in health care. The transmission of basic consumer information—taken for granted in virtually every other industry—is still a heroic task in the world of medicine. And the cost to patients in both convenience and safety is significant. Diagnoses are made and treatment selected without access to the patient’s entire medical history. Redundant testing due to missing information is accepted as a matter of course. Botched handoffs and lack of follow-through aggravate and jeopardize patients on a daily basis. Even the pace and breadth of medical research suffers because limited access to data sets hampers timely analysis.

At the root of these problems are ineffective business models and well-intentioned but now outdated regulations. Both need to change if the flow of medical information is to improve.

In health care most information technology (IT) companies continue to cling stubbornly to an outdated enterprise model. They charge enormous capital outlays for software acquisition and then levy exorbitant fees for ongoing service contracts. As long as the sale closes, they prosper. If their clients struggle with adoption or fall short on outcomes…they still prosper. Not only is there no incentive for information to flow freely across these systems but the companies selling them are profiting by helping to lock vital information within proprietary data silos.

Health care IT companies that fail to support the free flow of information may see a short-term business advantage, but any business model so divorced from social utility is destined for long-term failure. Much better to innovate, reinvent and evolve.

The fierce competitors in the financial services industry long ago concluded that they would all be much better off if information flowed. It is time for health care IT vendors to follow suit: compete on dimensions of service that benefit patients; create workflows that allow physicians to function at the top of their abilities; bring actionable information to caregivers at critical decision points; and support innovative care models. The CommonWell Health Alliance, a cross-industry collaboration advancing interoperability of patient data, is a promising first step in this direction.

Government also has a role to play. If data is to move seamlessly in health care, corporations, for-profit entities and entrepreneurs must have a financial incentive to make it move. Today, however, this is largely illegal. Outdated antikickback laws prevent health care organizations from charging for the secure transmission of medical information. These laws should be loosened to allow innovative businesses to build a sustainable market for the exchange of health information.

Finally, medical researchers need to shift efforts over time. Randomized clinical trials, still considered the gold standard of research, are too expensive and time-consuming to resolve more than a small proportion of important unanswered clinical questions. Often they focus on pristine, carefully selected populations to such an extent that results may not translate to more complex real-world conditions. While continuing with traditional clinical research methods, we must venture into the uncharted but promising terrain of analyzing huge quantities of electronic medical-record data. We need to place greater effort on studying individuals who receive the bulk of their care in community settings. New tools and statistical methodologies need to be developed to glean insight from data assets that are not designed to optimize research. Researchers need to reach out to organizations that have access to well-instrumented, real-time community medical data and demand it.

It will take some time but barriers to medical data interoperability will crumble—with profound benefits to society: Patient harm and inconvenience resulting from botched communications will become rare. More research will be conducted in real-world conditions, answering questions currently beyond reach. Consumers will be able to choose providers best positioned to address their concerns without struggling with the hassles of transferring information. And as medical information becomes seamlessly integrated with physician workflow, physicians can focus on what attracted them to medicine in the first place: caring for their patients.